Dan Lynch, James P. Gray, and Edward Rabinovitch, editors
SNA and TCP/IP Enterprise Networking
Manning Publications, 1997.
Pp. 540. $60.00. ISBN 0-13-127168-7

Reviewed by Daniel Lazenby <[email protected]>

Most organizations have their own share of local-area networks. Multiple networks (both local- and wide-area) using a variety of protocols may have been inherited as the result of a consolidation or merger. Whatever their origin, these networks need to play well together. Making them do so requires the network practitioner to make choices. These choices must maintain a balance among cost, reliability, achieving today's goals, remaining compatible with legacy architectures, and providing a migration path to yet-to-be-defined networking technologies. For many organizations, building a network over from the ground up is not an option.

SNA and TCP/IP Enterprise Networking, offered as a handbook for network practitioners, is based on the fundamental premise that multiple protocols will coexist forever. The authors feel this book will assist the reader with implementing reliable, cost-effective networking solutions with a migration path to future technologies.

The book first exposes the reader to a brief history of the development of multi-protocol communications. A brief discussion of the basic philosophical differences between SNA and TCP follows. The authors also describe important SNA multi-protocol integration products and popular deployment methods. The book closes with a look at emerging solutions. The authors illustrate both the strengths and weaknesses of each technology. While the book provides a balanced presentation, its primary view of the world is from an SNA perspective.

The TCP and SNA sections open with a set of functionally oriented tutorials. They introduce SNA terminology and acronyms, and present the what and why of a subject. Many of these tutorials walk you through the flow of an SNA process. A couple of these chapters introduce TCP/IP concepts and terminology. Other chapters present application programming interfaces (APIs) that may be used with either of the two network protocols.

Chapter 8, the last chapter of part I, summarizes the information presented in the first seven chapters. In many places it provides a side-by-side functional comparison of the two network architectures and clearly illustrates the similarities and differences in performance between them. The authors review the two protocols' usability and reliability characteristics, and they present approaches to and methods for converging the two protocols.

Following the tutorials is an examination of SNA's interoperability features. The book's second part, "SNA Interoperability Today," describes currently available solutions: multi-protocol routers, gateways, and other software. Part II opens with a description of today's popular SNA internetworking strategy's benefits and pitfalls. The next chapter examines why one might want to encapsulate SNA within TCP/IP and provides some alternatives to encapsulation. Another chapter presents an approach for using SNA as the transport for non-SNA protocols. Topics such as providing 3,270 terminals access to TCP/IP applications, and questions and considerations when selecting or implementing an SNA gateway, are covered in the last couple of chapters of part II.

Host (mainframe) systems and their SNA networks have not gone away. Some new mainframes are even being sold as "enterprise servers." Part II concludes with examining ways Web technology can be used to enhance, extend, and leverage legacy platforms and the SNA architecture. This platform-independent client/server technology is inexpensive, easy to use, and readily available for many platforms. Web browser technology may just turn out to be the great equalizer among platforms.

The third section of the book, "Emerging Solutions," discusses the impact of recent technology on SNA presentation services. It suggests that future SNA applications will be "Common User Access compliant client/server applications written in Java." In closing, the last chapter gazes into the protocol crystal ball, presenting one view of how distributed computing technology may generate the next revolution in protocols.

Overall, the authors offer a model for synthesizing a multi-vendor, multi-protocol network into a cohesive whole that can appear to applications and the user as a single integrated network.